Common Weakness Enumeration

CWE-501

Allowed

Trust Boundary Violation

Abstraction: Base · Status: Draft

The product mixes trusted and untrusted data in the same data structure or structured message.

58 vulnerabilities reference this CWE, most recent first.

GHSA-227W-WV4J-67H4

Vulnerability from github – Published: 2022-02-09 22:30 – Updated: 2026-01-22 20:53
VLAI
Summary
Class Loading Vulnerability in Artemis
Details

Impact

This affects all Artemis users who test Java assignments. Ares is not required. Students code that gets automatically tested can run arbitrary code in the container, or arbitrary code on the machine of an assessor in case of manual correction.

Patches

The problem cannot be resolved easily in Ares itself. Use the Maven Enforcer Plugin as follows:

<plugin>
    <groupId>org.apache.maven.plugins</groupId>
    <artifactId>maven-enforcer-plugin</artifactId>
    <version>3.0.0</version>
    <executions>
        <execution>
            <id>enforce-no-student-code-in-trusted-packages</id>
            <phase>process-classes</phase>
            <goals>
                <goal>enforce</goal>
            </goals>
        </execution>
    </executions>
    <configuration>
        <rules>
            <requireFilesDontExist>
                <files>
                    <!-- ADD HERE THE RULES ARES TELLS YOU ARE MISSING -->
                </files>
            </requireFilesDontExist>
        </rules>
    </configuration>
</plugin>

This fails the build if student classes reside in such packages that Ares trusts. Trusted packages added in Ares using @AddTrustedPackage should be added as well.

For more information

If you have any questions or comments about this advisory: * Open a discussion https://github.com/ls1intum/Ares/discussions * Open an issue in https://github.com/ls1intum/Ares/issues * Email us, see https://github.com/ls1intum/Ares/security/policy

References

See the assignment of Julius that passes the tests in TUM Artemis course: "Test - Praktikum: Grundlagen der Programmierung (Testkurs für Tutoren) - Security Tests" (if that still exists in 2022).

Also see #15 for almost the same problem.

Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "Maven",
        "name": "de.tum.in.ase:artemis-java-test-sandbox"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "1.8.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2024-23682"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-501",
      "CWE-653"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2022-02-09T22:30:30Z",
    "nvd_published_at": null,
    "severity": "HIGH"
  },
  "details": "### Impact\nThis affects all Artemis users who test Java assignments. **Ares is not required.**\nStudents code that gets automatically tested can run arbitrary code in the container,\nor arbitrary code on the machine of an assessor in case of manual correction.\n\n### Patches\nThe problem cannot be resolved easily in Ares itself. Use the Maven Enforcer Plugin as follows:\n\n```xml\n\u003cplugin\u003e\n    \u003cgroupId\u003eorg.apache.maven.plugins\u003c/groupId\u003e\n    \u003cartifactId\u003emaven-enforcer-plugin\u003c/artifactId\u003e\n    \u003cversion\u003e3.0.0\u003c/version\u003e\n    \u003cexecutions\u003e\n        \u003cexecution\u003e\n            \u003cid\u003eenforce-no-student-code-in-trusted-packages\u003c/id\u003e\n            \u003cphase\u003eprocess-classes\u003c/phase\u003e\n            \u003cgoals\u003e\n                \u003cgoal\u003eenforce\u003c/goal\u003e\n            \u003c/goals\u003e\n        \u003c/execution\u003e\n    \u003c/executions\u003e\n    \u003cconfiguration\u003e\n        \u003crules\u003e\n            \u003crequireFilesDontExist\u003e\n                \u003cfiles\u003e\n                    \u003c!-- ADD HERE THE RULES ARES TELLS YOU ARE MISSING --\u003e\n                \u003c/files\u003e\n            \u003c/requireFilesDontExist\u003e\n        \u003c/rules\u003e\n    \u003c/configuration\u003e\n\u003c/plugin\u003e\n```\n\nThis fails the build if student classes reside in such packages that Ares trusts. Trusted packages added in Ares using `@AddTrustedPackage` should be added as well.\n\n### For more information\nIf you have any questions or comments about this advisory:\n* Open a discussion https://github.com/ls1intum/Ares/discussions\n* Open an issue in https://github.com/ls1intum/Ares/issues\n* Email us, see https://github.com/ls1intum/Ares/security/policy\n\n### References\nSee the assignment of Julius that passes the tests in TUM Artemis course: \"Test - Praktikum: Grundlagen der Programmierung (Testkurs f\u00fcr Tutoren) - Security Tests\" (if that still exists in 2022).\n\nAlso see #15 for almost the same problem.",
  "id": "GHSA-227w-wv4j-67h4",
  "modified": "2026-01-22T20:53:41Z",
  "published": "2022-02-09T22:30:30Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/ls1intum/Ares/security/advisories/GHSA-227w-wv4j-67h4"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2024-23682"
    },
    {
      "type": "WEB",
      "url": "https://github.com/ls1intum/Ares/issues/15"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/ls1intum/Ares"
    },
    {
      "type": "WEB",
      "url": "https://github.com/ls1intum/Ares/releases/tag/1.8.0"
    },
    {
      "type": "WEB",
      "url": "https://vulncheck.com/advisories/vc-advisory-GHSA-227w-wv4j-67h4"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:R/S:C/C:H/I:H/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Class Loading Vulnerability in Artemis"
}

GHSA-22W6-C8H9-6MX7

Vulnerability from github – Published: 2026-03-31 18:31 – Updated: 2026-03-31 18:31
VLAI
Details

NVIDIA Jetson Linux has a vulnerability in initrd, where the nvluks trusted application is not disabled. A successful exploit of this vulnerability might lead to information disclosure.

Show details on source website

{
  "affected": [],
  "aliases": [
    "CVE-2026-24153"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-501"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2026-03-31T17:16:30Z",
    "severity": "MODERATE"
  },
  "details": "NVIDIA Jetson Linux has a vulnerability in initrd, where the nvluks trusted application is not disabled. A successful exploit of this vulnerability might lead to information disclosure.",
  "id": "GHSA-22w6-c8h9-6mx7",
  "modified": "2026-03-31T18:31:31Z",
  "published": "2026-03-31T18:31:31Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-24153"
    },
    {
      "type": "WEB",
      "url": "https://nvidia.custhelp.com/app/answers/detail/a_id/5797"
    },
    {
      "type": "WEB",
      "url": "https://www.cve.org/CVERecord?id=CVE-2026-24153"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:P/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-36JR-8W83-WR8Q

Vulnerability from github – Published: 2024-11-12 18:30 – Updated: 2024-11-12 18:30
VLAI
Details

Visual Studio Code Python Extension Remote Code Execution Vulnerability

Show details on source website

{
  "affected": [],
  "aliases": [
    "CVE-2024-49050"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-501"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2024-11-12T18:15:45Z",
    "severity": "HIGH"
  },
  "details": "Visual Studio Code Python Extension Remote Code Execution Vulnerability",
  "id": "GHSA-36jr-8w83-wr8q",
  "modified": "2024-11-12T18:30:59Z",
  "published": "2024-11-12T18:30:59Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2024-49050"
    },
    {
      "type": "WEB",
      "url": "https://msrc.microsoft.com/update-guide/vulnerability/CVE-2024-49050"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-5R25-P9F2-W2XV

Vulnerability from github – Published: 2025-02-19 18:32 – Updated: 2026-03-25 00:31
VLAI
Details

A flaw was found in grub2. Grub's dump command is not blocked when grub is in lockdown mode, which allows the user to read any memory information, and an attacker may leverage this in order to extract signatures, salts, and other sensitive information from the memory.

Show details on source website

{
  "affected": [],
  "aliases": [
    "CVE-2025-1118"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-501"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2025-02-19T18:15:24Z",
    "severity": "MODERATE"
  },
  "details": "A flaw was found in grub2. Grub\u0027s dump command is not blocked when grub is in lockdown mode, which allows the user to read any memory information, and an attacker may leverage this in order to extract signatures, salts, and other sensitive information from the memory.",
  "id": "GHSA-5r25-p9f2-w2xv",
  "modified": "2026-03-25T00:31:11Z",
  "published": "2025-02-19T18:32:24Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-1118"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2025:16154"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/security/cve/CVE-2025-1118"
    },
    {
      "type": "WEB",
      "url": "https://bugzilla.redhat.com/show_bug.cgi?id=2346137"
    },
    {
      "type": "WEB",
      "url": "https://git.savannah.gnu.org/cgit/grub.git/commit/?id=34824806ac6302f91e8cabaa41308eaced25725f"
    },
    {
      "type": "WEB",
      "url": "https://lists.gnu.org/archive/html/grub-devel/2025-02/msg00024.html"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:L/AC:L/PR:H/UI:N/S:U/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-6VRV-94JV-CRRG

Vulnerability from github – Published: 2020-07-07 00:01 – Updated: 2021-01-07 23:48
VLAI
Summary
Context isolation bypass via Promise in Electron
Details

Impact

Apps using contextIsolation are affected.

This is a context isolation bypass, meaning that code running in the main world context in the renderer can reach into the isolated Electron context and perform privileged actions.

Workarounds

There are no app-side workarounds, you must update your Electron version to be protected.

Fixed Versions

  • 9.0.0-beta.21
  • 8.2.4
  • 7.2.4
  • 6.1.11

For more information

If you have any questions or comments about this advisory: * Email us at security@electronjs.org

Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "npm",
        "name": "electron"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "6.1.11"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    },
    {
      "package": {
        "ecosystem": "npm",
        "name": "electron"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "7.0.0"
            },
            {
              "fixed": "7.2.4"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    },
    {
      "package": {
        "ecosystem": "npm",
        "name": "electron"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "8.0.0"
            },
            {
              "fixed": "8.2.4"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2020-15096"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-501"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2020-07-06T23:55:38Z",
    "nvd_published_at": null,
    "severity": "LOW"
  },
  "details": "### Impact\nApps using `contextIsolation` are affected.\n\nThis is a context isolation bypass, meaning that code running in the main world context in the renderer can reach into the isolated Electron context and perform privileged actions.\n\n### Workarounds\nThere are no app-side workarounds, you must update your Electron version to be protected.\n\n### Fixed Versions\n* `9.0.0-beta.21`\n* `8.2.4`\n* `7.2.4`\n* `6.1.11`\n\n### For more information\nIf you have any questions or comments about this advisory:\n* Email us at [security@electronjs.org](mailto:security@electronjs.org)",
  "id": "GHSA-6vrv-94jv-crrg",
  "modified": "2021-01-07T23:48:19Z",
  "published": "2020-07-07T00:01:05Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/electron/electron/security/advisories/GHSA-6vrv-94jv-crrg"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2020-15096"
    },
    {
      "type": "WEB",
      "url": "https://www.electronjs.org/releases/stable?page=3#release-notes-for-v824"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:C/C:N/I:H/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Context isolation bypass via Promise in Electron"
}

GHSA-7328-G372-24VF

Vulnerability from github – Published: 2026-01-13 15:37 – Updated: 2026-01-15 12:30
VLAI
Details

Incorrect boundary conditions in the Graphics component. This vulnerability affects Firefox < 147, Firefox ESR < 115.32, and Firefox ESR < 140.7.

Show details on source website

{
  "affected": [],
  "aliases": [
    "CVE-2026-0886"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-119",
      "CWE-501"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2026-01-13T14:16:39Z",
    "severity": "MODERATE"
  },
  "details": "Incorrect boundary conditions in the Graphics component. This vulnerability affects Firefox \u003c 147, Firefox ESR \u003c 115.32, and Firefox ESR \u003c 140.7.",
  "id": "GHSA-7328-g372-24vf",
  "modified": "2026-01-15T12:30:26Z",
  "published": "2026-01-13T15:37:04Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-0886"
    },
    {
      "type": "WEB",
      "url": "https://bugzilla.mozilla.org/show_bug.cgi?id=2005658"
    },
    {
      "type": "WEB",
      "url": "https://www.mozilla.org/security/advisories/mfsa2026-01"
    },
    {
      "type": "WEB",
      "url": "https://www.mozilla.org/security/advisories/mfsa2026-02"
    },
    {
      "type": "WEB",
      "url": "https://www.mozilla.org/security/advisories/mfsa2026-03"
    },
    {
      "type": "WEB",
      "url": "https://www.mozilla.org/security/advisories/mfsa2026-04"
    },
    {
      "type": "WEB",
      "url": "https://www.mozilla.org/security/advisories/mfsa2026-05"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-75MJ-4G74-9RG2

Vulnerability from github – Published: 2025-12-13 18:30 – Updated: 2025-12-15 23:55
VLAI
Summary
Universal Tool Calling Protocol (UTCP) client library for Python vulnerable to Trust Boundary Violation through Manual JSON specification
Details

The vulnerability arises when a client fetches a tools’ JSON specification, known as a Manual, from a remote Manual Endpoint. While a provider may initially serve a benign manual (e.g., one defining an HTTP tool call), earning the clients’ trust, a malicious provider can later change the manual to exploit the client.

Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "utcp"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "1.1.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2025-14542"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-501"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2025-12-15T23:55:42Z",
    "nvd_published_at": "2025-12-13T16:16:51Z",
    "severity": "HIGH"
  },
  "details": "The vulnerability arises when a client fetches a tools\u2019 JSON specification, known as a Manual, from a remote Manual Endpoint. While a provider may initially serve a benign manual (e.g., one defining an HTTP tool call), earning the clients\u2019 trust, a malicious provider can later change the manual to exploit the client.",
  "id": "GHSA-75mj-4g74-9rg2",
  "modified": "2025-12-15T23:55:42Z",
  "published": "2025-12-13T18:30:21Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-14542"
    },
    {
      "type": "WEB",
      "url": "https://github.com/universal-tool-calling-protocol/python-utcp/commit/2dc9c02df72cad3770c934959325ec344b441444"
    },
    {
      "type": "WEB",
      "url": "https://github.com/universal-tool-calling-protocol/python-utcp"
    },
    {
      "type": "WEB",
      "url": "https://research.jfrog.com/vulnerabilities/python-utcp-untrusted-manual-command-execution-jfsa-2025-001648329"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Universal Tool Calling Protocol (UTCP) client library for Python vulnerable to Trust Boundary Violation through Manual JSON specification"
}

GHSA-76MC-F452-CXCM

Vulnerability from github – Published: 2026-06-15 19:59 – Updated: 2026-07-23 14:01
VLAI
Summary
DOMPurify: Hook mutation of `data.allowedTags` / `data.allowedAttributes` permanently pollutes `DEFAULT_ALLOWED_TAGS` / `DEFAULT_ALLOWED_ATTR`
Details

Hook mutation of data.allowedTags / data.allowedAttributes permanently pollutes DEFAULT_ALLOWED_TAGS / DEFAULT_ALLOWED_ATTR

CWE: CWE-501 (Trust Boundary Violation — hook-scoped mutation leaks to global default sets) via CWE-693 (Protection Mechanism Failure — the default allow-list is silently widened for all subsequent sanitize calls)

Summary

The data.allowedTags and data.allowedAttributes fields passed to uponSanitizeElement and uponSanitizeAttribute hooks are direct references to the library's live ALLOWED_TAGS / ALLOWED_ATTR sets. For sanitize calls that don't supply an explicit cfg.ALLOWED_TAGS / cfg.ALLOWED_ATTR array, those live sets are themselves direct references to the module-level DEFAULT_ALLOWED_TAGS / DEFAULT_ALLOWED_ATTR constants. A hook that mutates these fields — a natural-looking pattern for "allow X for this iteration" — permanently writes new entries into the default constants for the DOMPurify instance's lifetime. Every subsequent sanitize call that doesn't override the config inherits the widened defaults, so an attacker payload that uses the poisoned tag/attribute name survives sanitization. removeAllHooks(), clearConfig(), and even passing a fresh cfg: {} do not recover; only constructing a new DOMPurify instance does.

The maintainer's existing defense at src/purify.ts:696-700 explicitly clones DEFAULT_ALLOWED_TAGS before mutating it via cfg.ADD_TAGS (array form), demonstrating awareness of this exact class. The hook path remained uncovered.

Affected

  • DOMPurify ≤ 3.4.5, including main at 7996f1dc78eb8b7922388aed75d94a9f8fad9a36
  • Any application that installs a hook on uponSanitizeElement or uponSanitizeAttribute that writes to data.allowedTags[...] = true or data.allowedAttributes[...] = true and later sanitizes attacker-influenced content with default config (no explicit cfg.ALLOWED_TAGS / cfg.ALLOWED_ATTR array)

Vulnerability details

[A] — data.allowedTags is a reference to ALLOWED_TAGS

src/purify.ts:1206-1209:

_executeHooks(hooks.uponSanitizeElement, currentNode, {
  tagName,
  allowedTags: ALLOWED_TAGS,         // [A] direct reference; hook mutation
                                      //     mutates the very ALLOWED_TAGS the
                                      //     library checks on the next element
});

src/purify.ts:1494-1500 (the matching attribute hook):

const hookEvent = {
  attrName: '',
  attrValue: '',
  keepAttr: true,
  allowedAttributes: ALLOWED_ATTR,    // [A'] same pattern
  forceKeepAttr: undefined,
};

[B] — ALLOWED_TAGS = DEFAULT_ALLOWED_TAGS for default-cfg sanitize calls

src/purify.ts:527-531:

ALLOWED_TAGS =
  objectHasOwnProperty(cfg, 'ALLOWED_TAGS') &&
  arrayIsArray(cfg.ALLOWED_TAGS)
    ? addToSet({}, cfg.ALLOWED_TAGS, transformCaseFunc)
    : DEFAULT_ALLOWED_TAGS;            // [B] reference assignment; ALLOWED_TAGS
                                       //     IS the DEFAULT_ALLOWED_TAGS object

(The ALLOWED_ATTR = DEFAULT_ALLOWED_ATTR path at :532-536 is symmetric.)

The mismatch

A hook author who writes data.allowedTags['script'] = true reasonably expects per-call scope — the API name is "data", suggesting per-event payload. But [A] makes this a direct reference, and [B] makes that reference equal to the module-level default for the common default-cfg path. The hook's mutation therefore writes to a constant that every subsequent default-cfg sanitize call rebinds to.

The maintainer already recognized this class for the ADD_TAGS array path — src/purify.ts:696-700:

} else if (arrayIsArray(cfg.ADD_TAGS)) {
  if (ALLOWED_TAGS === DEFAULT_ALLOWED_TAGS) {
    ALLOWED_TAGS = clone(ALLOWED_TAGS);   // explicitly clone DEFAULT before
                                          // mutating to avoid this pollution
  }
  addToSet(ALLOWED_TAGS, cfg.ADD_TAGS, transformCaseFunc);
}

The same defensive clone is missing from the hook code paths.

Proof of concept

// 1) fresh DOMPurify, default config — script is blocked
DOMPurify.sanitize('<svg><script>alert(1)</script></svg>');
// → "<svg></svg>"

// 2) install a hook that mutates data.allowedTags (natural-looking pattern)
DOMPurify.addHook('uponSanitizeElement', (node, data) => {
  data.allowedTags['script'] = true;
});

// 3) one sanitize call WITH the hook — script survives (expected during the hook)
DOMPurify.sanitize('<svg><script>alert(1)</script></svg>');
// → "<svg><script>alert(1)</script></svg>"

// 4) remove the hook
DOMPurify.removeAllHooks();
DOMPurify.clearConfig();

// 5) sanitize attacker content with default config — POLLUTION PERSISTS
DOMPurify.sanitize('<svg><script>alert(1)</script></svg>');
// → "<svg><script>alert(1)</script></svg>"  ← script survived without any hook

// 6) the only recovery: create a fresh DOMPurify instance
const fresh = DOMPurify(window);
fresh.sanitize('<svg><script>alert(1)</script></svg>');
// → "<svg></svg>"  ← clean

Observed (Chromium 148.0.7778.96, DOMPurify HEAD 7996f1d):

step input output bypass?
1 fresh baseline <svg><script>__</script></svg> <svg></svg> no
1b fresh baseline <a onclick=__>x</a> <a>x</a> no
2 with hook (script) <svg><script>__</script></svg> <svg><script>__</script></svg> yes (expected)
2b with hook (onclick) <a onclick=__>x</a> <a onclick="__">x</a> yes (expected)
3 after removeAllHooks() same <svg><script>__</script></svg> YES (pollution)
3b after removeAllHooks() same <a onclick="__">x</a> YES (pollution)
4 after clearConfig() same <svg><script>__</script></svg> YES
4b after clearConfig() same <a onclick="__">x</a> YES
5 explicit restrictive cfg.ALLOWED_TAGS=['svg'] same <svg></svg> no (cloned set)
6 back to no cfg same <svg><script>__</script></svg> YES
6b back to no cfg same <a onclick="__">x</a> YES
7 fresh DOMPurify(window) instance same <svg></svg> no
7b fresh instance <a onclick=__>x</a> <a>x</a> no

Impact

Direct

Any application using DOMPurify that has any registered hook with the pattern data.allowedTags[...] = true or data.allowedAttributes[...] = true. The hook need not be designed to be permissive — it might be intended to temporarily allow a custom tag for one specific element shape. After the hook has executed even once, every subsequent default-config sanitize call carries the widened defaults, including:

  • attacker content rendered via separate code paths (e.g., the same library serving a comments section and a profile bio, where the bio uses the hook and the comments use plain DOMPurify.sanitize(text))
  • third-party libraries that call DOMPurify.sanitize on the same instance

The bypass survives DOMPurify.removeAllHooks() and DOMPurify.clearConfig() — the obvious "reset" calls a dev would reach for. Detection requires reading the DEFAULT_ALLOWED_TAGS / DEFAULT_ALLOWED_ATTR sets directly, which are not part of the public API.

Indirect / second-order

  • Editor / preview libraries that compose with DOMPurify — if any consumer registers a hook that mutates data.allowedTags, every other consumer's sanitize calls inherit the widening.
  • Test suites that exercise multiple sanitize configurations — once a test's hook pollutes the defaults, later tests that assume default behavior may pass with widened defaults and miss real regressions.
  • Long-running servers (SSR, edge functions) that reuse a single DOMPurify instance — pollution accumulates over the process lifetime.

Why the existing maintainer defense for ADD_TAGS doesn't catch this

src/purify.ts:696-700 already documents awareness:

} else if (arrayIsArray(cfg.ADD_TAGS)) {
  if (ALLOWED_TAGS === DEFAULT_ALLOWED_TAGS) {
    ALLOWED_TAGS = clone(ALLOWED_TAGS);
  }
  addToSet(ALLOWED_TAGS, cfg.ADD_TAGS, transformCaseFunc);
}

The clone-before-mutate pattern is exactly what's needed at the hook callsites (:1206-1209 and :1494-1500) but was not extended there. The new entries this report's bypass adds to the defaults survive the same way ADD_TAGS array entries would have survived before that fix landed.

Suggested fix

Three minimal-impact options, in order of preference:

  1. Hand the hook a defensive copy (most surgical):

ts _executeHooks(hooks.uponSanitizeElement, currentNode, { tagName, allowedTags: { ...ALLOWED_TAGS }, // shallow copy; mutations stay scoped });

Doc note: "data.allowedTags is a snapshot; to widen the live set, use cfg.ADD_TAGS or set the value to true in the snapshot and check the snapshot from a subsequent attribute hook." Hooks that read it for inspection still work; hooks that intended cross-call mutation must be rewritten to use a proper config path (which is the correct API anyway).

  1. Clone-on-write inside the hook path, mirroring the existing ADD_TAGS defense at :696-700: detect that ALLOWED_TAGS === DEFAULT_ALLOWED_TAGS after the hook returns, and if so, replace it with a clone for subsequent processing. This preserves the live-mutation semantics for in-call effects while preventing cross-call leakage.

  2. Lazy-clone ALLOWED_TAGS/ALLOWED_ATTR from defaults on first mutation: install a Proxy or accessor that triggers a clone before mutation. Largest surface area, but bulletproof.

Option (1) is the cleanest API contract: hook event objects should be event-local, never references to library-internal state.

Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "npm",
        "name": "dompurify"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "3.4.7"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-65902"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-501",
      "CWE-693"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-15T19:59:09Z",
    "nvd_published_at": null,
    "severity": "MODERATE"
  },
  "details": "# Hook mutation of `data.allowedTags` / `data.allowedAttributes` permanently pollutes `DEFAULT_ALLOWED_TAGS` / `DEFAULT_ALLOWED_ATTR`\n\n**CWE**: CWE-501 (Trust Boundary Violation \u2014 hook-scoped mutation leaks to global default sets) via CWE-693 (Protection Mechanism Failure \u2014 the default allow-list is silently widened for all subsequent sanitize calls)\n\n## Summary\n\nThe `data.allowedTags` and `data.allowedAttributes` fields passed to `uponSanitizeElement` and `uponSanitizeAttribute` hooks are **direct references** to the library\u0027s live `ALLOWED_TAGS` / `ALLOWED_ATTR` sets. For sanitize calls that don\u0027t supply an explicit `cfg.ALLOWED_TAGS` / `cfg.ALLOWED_ATTR` array, those live sets are themselves direct references to the module-level `DEFAULT_ALLOWED_TAGS` / `DEFAULT_ALLOWED_ATTR` constants. A hook that mutates these fields \u2014 a natural-looking pattern for \"allow `X` for this iteration\" \u2014 permanently writes new entries into the default constants for the DOMPurify instance\u0027s lifetime. Every subsequent sanitize call that doesn\u0027t override the config inherits the widened defaults, so an attacker payload that uses the poisoned tag/attribute name survives sanitization. `removeAllHooks()`, `clearConfig()`, and even passing a fresh `cfg: {}` do not recover; only constructing a new DOMPurify instance does.\n\nThe maintainer\u0027s existing defense at `src/purify.ts:696-700` explicitly clones `DEFAULT_ALLOWED_TAGS` before mutating it via `cfg.ADD_TAGS` (array form), demonstrating awareness of this exact class. The hook path remained uncovered.\n\n## Affected\n\n- DOMPurify \u2264 3.4.5, including `main` at `7996f1dc78eb8b7922388aed75d94a9f8fad9a36`\n- Any application that installs a hook on `uponSanitizeElement` or `uponSanitizeAttribute` that writes to `data.allowedTags[...] = true` or `data.allowedAttributes[...] = true` and later sanitizes attacker-influenced content with default config (no explicit `cfg.ALLOWED_TAGS` / `cfg.ALLOWED_ATTR` array)\n\n## Vulnerability details\n\n### [A] \u2014 `data.allowedTags` is a reference to `ALLOWED_TAGS`\n\n`src/purify.ts:1206-1209`:\n\n```ts\n_executeHooks(hooks.uponSanitizeElement, currentNode, {\n  tagName,\n  allowedTags: ALLOWED_TAGS,         // [A] direct reference; hook mutation\n                                      //     mutates the very ALLOWED_TAGS the\n                                      //     library checks on the next element\n});\n```\n\n`src/purify.ts:1494-1500` (the matching attribute hook):\n\n```ts\nconst hookEvent = {\n  attrName: \u0027\u0027,\n  attrValue: \u0027\u0027,\n  keepAttr: true,\n  allowedAttributes: ALLOWED_ATTR,    // [A\u0027] same pattern\n  forceKeepAttr: undefined,\n};\n```\n\n### [B] \u2014 `ALLOWED_TAGS = DEFAULT_ALLOWED_TAGS` for default-cfg sanitize calls\n\n`src/purify.ts:527-531`:\n\n```ts\nALLOWED_TAGS =\n  objectHasOwnProperty(cfg, \u0027ALLOWED_TAGS\u0027) \u0026\u0026\n  arrayIsArray(cfg.ALLOWED_TAGS)\n    ? addToSet({}, cfg.ALLOWED_TAGS, transformCaseFunc)\n    : DEFAULT_ALLOWED_TAGS;            // [B] reference assignment; ALLOWED_TAGS\n                                       //     IS the DEFAULT_ALLOWED_TAGS object\n```\n\n(The `ALLOWED_ATTR = DEFAULT_ALLOWED_ATTR` path at `:532-536` is symmetric.)\n\n### The mismatch\n\nA hook author who writes `data.allowedTags[\u0027script\u0027] = true` reasonably expects per-call scope \u2014 the API name is *\"data\"*, suggesting per-event payload. But [A] makes this a direct reference, and [B] makes that reference equal to the module-level default for the common default-cfg path. The hook\u0027s mutation therefore writes to a *constant* that every subsequent default-cfg sanitize call rebinds to.\n\nThe maintainer already recognized this class for the `ADD_TAGS` array path \u2014 `src/purify.ts:696-700`:\n\n```ts\n} else if (arrayIsArray(cfg.ADD_TAGS)) {\n  if (ALLOWED_TAGS === DEFAULT_ALLOWED_TAGS) {\n    ALLOWED_TAGS = clone(ALLOWED_TAGS);   // explicitly clone DEFAULT before\n                                          // mutating to avoid this pollution\n  }\n  addToSet(ALLOWED_TAGS, cfg.ADD_TAGS, transformCaseFunc);\n}\n```\n\nThe same defensive clone is missing from the hook code paths.\n\n## Proof of concept\n\n```js\n// 1) fresh DOMPurify, default config \u2014 script is blocked\nDOMPurify.sanitize(\u0027\u003csvg\u003e\u003cscript\u003ealert(1)\u003c/script\u003e\u003c/svg\u003e\u0027);\n// \u2192 \"\u003csvg\u003e\u003c/svg\u003e\"\n\n// 2) install a hook that mutates data.allowedTags (natural-looking pattern)\nDOMPurify.addHook(\u0027uponSanitizeElement\u0027, (node, data) =\u003e {\n  data.allowedTags[\u0027script\u0027] = true;\n});\n\n// 3) one sanitize call WITH the hook \u2014 script survives (expected during the hook)\nDOMPurify.sanitize(\u0027\u003csvg\u003e\u003cscript\u003ealert(1)\u003c/script\u003e\u003c/svg\u003e\u0027);\n// \u2192 \"\u003csvg\u003e\u003cscript\u003ealert(1)\u003c/script\u003e\u003c/svg\u003e\"\n\n// 4) remove the hook\nDOMPurify.removeAllHooks();\nDOMPurify.clearConfig();\n\n// 5) sanitize attacker content with default config \u2014 POLLUTION PERSISTS\nDOMPurify.sanitize(\u0027\u003csvg\u003e\u003cscript\u003ealert(1)\u003c/script\u003e\u003c/svg\u003e\u0027);\n// \u2192 \"\u003csvg\u003e\u003cscript\u003ealert(1)\u003c/script\u003e\u003c/svg\u003e\"  \u2190 script survived without any hook\n\n// 6) the only recovery: create a fresh DOMPurify instance\nconst fresh = DOMPurify(window);\nfresh.sanitize(\u0027\u003csvg\u003e\u003cscript\u003ealert(1)\u003c/script\u003e\u003c/svg\u003e\u0027);\n// \u2192 \"\u003csvg\u003e\u003c/svg\u003e\"  \u2190 clean\n```\n\nObserved (Chromium 148.0.7778.96, DOMPurify HEAD `7996f1d`):\n\n| step | input | output | bypass? |\n|---|---|---|---|\n| 1 fresh baseline | `\u003csvg\u003e\u003cscript\u003e__\u003c/script\u003e\u003c/svg\u003e` | `\u003csvg\u003e\u003c/svg\u003e` | no |\n| 1b fresh baseline | `\u003ca onclick=__\u003ex\u003c/a\u003e` | `\u003ca\u003ex\u003c/a\u003e` | no |\n| 2 with hook (script) | `\u003csvg\u003e\u003cscript\u003e__\u003c/script\u003e\u003c/svg\u003e` | `\u003csvg\u003e\u003cscript\u003e__\u003c/script\u003e\u003c/svg\u003e` | yes (expected) |\n| 2b with hook (onclick) | `\u003ca onclick=__\u003ex\u003c/a\u003e` | `\u003ca onclick=\"__\"\u003ex\u003c/a\u003e` | yes (expected) |\n| 3 after `removeAllHooks()` | same | `\u003csvg\u003e\u003cscript\u003e__\u003c/script\u003e\u003c/svg\u003e` | **YES (pollution)** |\n| 3b after `removeAllHooks()` | same | `\u003ca onclick=\"__\"\u003ex\u003c/a\u003e` | **YES (pollution)** |\n| 4 after `clearConfig()` | same | `\u003csvg\u003e\u003cscript\u003e__\u003c/script\u003e\u003c/svg\u003e` | **YES** |\n| 4b after `clearConfig()` | same | `\u003ca onclick=\"__\"\u003ex\u003c/a\u003e` | **YES** |\n| 5 explicit restrictive `cfg.ALLOWED_TAGS=[\u0027svg\u0027]` | same | `\u003csvg\u003e\u003c/svg\u003e` | no (cloned set) |\n| 6 back to no cfg | same | `\u003csvg\u003e\u003cscript\u003e__\u003c/script\u003e\u003c/svg\u003e` | **YES** |\n| 6b back to no cfg | same | `\u003ca onclick=\"__\"\u003ex\u003c/a\u003e` | **YES** |\n| 7 fresh `DOMPurify(window)` instance | same | `\u003csvg\u003e\u003c/svg\u003e` | no |\n| 7b fresh instance | `\u003ca onclick=__\u003ex\u003c/a\u003e` | `\u003ca\u003ex\u003c/a\u003e` | no |\n\n## Impact\n\n### Direct\n\nAny application using `DOMPurify` that has any registered hook with the pattern `data.allowedTags[...] = true` or `data.allowedAttributes[...] = true`. The hook need not be designed to be permissive \u2014 it might be intended to *temporarily* allow a custom tag for one specific element shape. After the hook has executed even once, every subsequent default-config sanitize call carries the widened defaults, including:\n\n- attacker content rendered via separate code paths (e.g., the same library serving a comments section and a profile bio, where the bio uses the hook and the comments use plain `DOMPurify.sanitize(text)`)\n- third-party libraries that call `DOMPurify.sanitize` on the same instance\n\nThe bypass survives `DOMPurify.removeAllHooks()` and `DOMPurify.clearConfig()` \u2014 the obvious \"reset\" calls a dev would reach for. Detection requires reading the `DEFAULT_ALLOWED_TAGS` / `DEFAULT_ALLOWED_ATTR` sets directly, which are not part of the public API.\n\n### Indirect / second-order\n\n- **Editor / preview libraries** that compose with DOMPurify \u2014 if any consumer registers a hook that mutates `data.allowedTags`, every other consumer\u0027s sanitize calls inherit the widening.\n- **Test suites** that exercise multiple sanitize configurations \u2014 once a test\u0027s hook pollutes the defaults, later tests that assume default behavior may pass with widened defaults and miss real regressions.\n- **Long-running servers** (SSR, edge functions) that reuse a single DOMPurify instance \u2014 pollution accumulates over the process lifetime.\n\n### Why the existing maintainer defense for `ADD_TAGS` doesn\u0027t catch this\n\n`src/purify.ts:696-700` already documents awareness:\n\n```ts\n} else if (arrayIsArray(cfg.ADD_TAGS)) {\n  if (ALLOWED_TAGS === DEFAULT_ALLOWED_TAGS) {\n    ALLOWED_TAGS = clone(ALLOWED_TAGS);\n  }\n  addToSet(ALLOWED_TAGS, cfg.ADD_TAGS, transformCaseFunc);\n}\n```\n\nThe clone-before-mutate pattern is exactly what\u0027s needed at the hook callsites (`:1206-1209` and `:1494-1500`) but was not extended there. The new entries this report\u0027s bypass adds to the defaults survive the same way `ADD_TAGS` array entries would have survived before that fix landed.\n\n## Suggested fix\n\nThree minimal-impact options, in order of preference:\n\n1. **Hand the hook a defensive copy** (most surgical):\n\n   ```ts\n   _executeHooks(hooks.uponSanitizeElement, currentNode, {\n     tagName,\n     allowedTags: { ...ALLOWED_TAGS },     // shallow copy; mutations stay scoped\n   });\n   ```\n\n   Doc note: \"`data.allowedTags` is a snapshot; to widen the live set, use `cfg.ADD_TAGS` or set the value to true in the snapshot and check the snapshot from a subsequent attribute hook.\" Hooks that read it for inspection still work; hooks that intended cross-call mutation must be rewritten to use a proper config path (which is the correct API anyway).\n\n2. **Clone-on-write inside the hook path**, mirroring the existing `ADD_TAGS` defense at `:696-700`: detect that `ALLOWED_TAGS === DEFAULT_ALLOWED_TAGS` after the hook returns, and if so, replace it with a clone for subsequent processing. This preserves the live-mutation semantics for in-call effects while preventing cross-call leakage.\n\n3. **Lazy-clone `ALLOWED_TAGS`/`ALLOWED_ATTR` from defaults on first mutation**: install a Proxy or accessor that triggers a clone before mutation. Largest surface area, but bulletproof.\n\nOption (1) is the cleanest API contract: hook event objects should be event-local, never references to library-internal state.",
  "id": "GHSA-76mc-f452-cxcm",
  "modified": "2026-07-23T14:01:17Z",
  "published": "2026-06-15T19:59:09Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/cure53/DOMPurify/security/advisories/GHSA-76mc-f452-cxcm"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/cure53/DOMPurify"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "DOMPurify: Hook mutation of `data.allowedTags` / `data.allowedAttributes` permanently pollutes `DEFAULT_ALLOWED_TAGS` / `DEFAULT_ALLOWED_ATTR`"
}

GHSA-7V95-W7C6-3W2G

Vulnerability from github – Published: 2023-09-25 18:30 – Updated: 2024-04-04 07:50
VLAI
Details

Docker Desktop 4.11.x allows --no-windows-containers flag bypass via IPC response spoofing which may lead to Local Privilege Escalation (LPE).This issue affects Docker Desktop: 4.11.X.

Show details on source website

{
  "affected": [],
  "aliases": [
    "CVE-2023-0627"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-501"
    ],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2023-09-25T16:15:13Z",
    "severity": "HIGH"
  },
  "details": "Docker Desktop 4.11.x allows --no-windows-containers flag bypass via IPC response spoofing which may lead to Local Privilege Escalation (LPE).This issue affects Docker Desktop: 4.11.X.\n\n",
  "id": "GHSA-7v95-w7c6-3w2g",
  "modified": "2024-04-04T07:50:11Z",
  "published": "2023-09-25T18:30:50Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2023-0627"
    },
    {
      "type": "WEB",
      "url": "https://docs.docker.com/desktop/release-notes/#4120"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
      "type": "CVSS_V3"
    }
  ]
}

GHSA-94F4-HR76-P5J6

Vulnerability from github – Published: 2026-06-16 17:36 – Updated: 2026-09-01 15:30
VLAI
Summary
vLLM: OpenAI auth bypass
Details

Summary

A vulnerability in ASGI web servers and starlette's trust on those web servers enables an authentication bypass of the OpenAI API AuthenticationMiddleware, which was discovered during @x41sec's source code audit. It allows to use the API without providing the configured VLLM_API_KEY or --api-key.

Details

In https://github.com/vllm-project/vllm/blob/v0.14.0/vllm/entrypoints/openai/api_server.py#L689-L692 the url_path is taken from the URL, which is reconstructed by starlette based on the request scope.

from starlette.datastructures import URL, Headers, MutableHeaders, State

# ...

url_path = URL(scope=scope).path.removeprefix(root_path)
headers = Headers(scope=scope)
if url_path.startswith("/v1") and not self.verify_token(headers):
    response = JSONResponse(content={"error": "Unauthorized"}, status_code=401)
    return response(scope, receive, send)
return self.app(scope, receive, send)

The request scope includes the request's Host: header and reconstructs the URL as shown below:

f"{scheme}://{host_header}{path}"

Neither starlette nor any of the ASGI servers (including uvicorn, which vllm uses) properly filter the Host: header for invalid characters. This allows an attacker to include special URL characters such as / or ? in the Host: header and thereby control the reconstructed URL and it's .path attribute.

FastAPI/starlette's routing uses the HTTP path and does not depend on the parsed url.path attribute, allowing attackers to reach an endpoint via a certain path while providing a different value in the .path.

Impact

  • Instances of vllm that use an API Key for the OpenAI API and expose the API to attackers.
  • Instances behind an RFC-conforming web server (such as nginx) are not affected.
Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "PyPI",
        "name": "vllm"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0.3.0"
            },
            {
              "fixed": "0.22.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-48746"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-444",
      "CWE-501"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-06-16T17:36:41Z",
    "nvd_published_at": "2026-06-22T23:16:30Z",
    "severity": "CRITICAL"
  },
  "details": "### Summary\n\nA vulnerability in ASGI web servers and starlette\u0027s trust on those web servers enables an authentication bypass of the OpenAI API `AuthenticationMiddleware`, which was discovered during @x41sec\u0027s source code audit.\nIt allows to use the API without providing the configured `VLLM_API_KEY` or `--api-key`.\n\n### Details\n\nIn https://github.com/vllm-project/vllm/blob/v0.14.0/vllm/entrypoints/openai/api_server.py#L689-L692 the `url_path` is taken from the `URL`, which is reconstructed by _starlette_ based on the request `scope`.\n\n```py\nfrom starlette.datastructures import URL, Headers, MutableHeaders, State\n\n# ...\n\nurl_path = URL(scope=scope).path.removeprefix(root_path)\nheaders = Headers(scope=scope)\nif url_path.startswith(\"/v1\") and not self.verify_token(headers):\n    response = JSONResponse(content={\"error\": \"Unauthorized\"}, status_code=401)\n    return response(scope, receive, send)\nreturn self.app(scope, receive, send)\n```\n\nThe request `scope` includes the request\u0027s `Host:` header and reconstructs the URL as shown below:\n\n```py\nf\"{scheme}://{host_header}{path}\"\n```\n\nNeither starlette nor [any of the ASGI servers](https://asgi.readthedocs.io/en/latest/implementations.html#servers) (including uvicorn, which vllm uses) properly filter the `Host:` header for invalid characters. This allows an attacker to include special URL characters such as `/` or `?` in the `Host:` header and thereby control the reconstructed URL and it\u0027s `.path` attribute.\n\nFastAPI/starlette\u0027s routing uses the HTTP path and does not depend on the parsed url.path attribute, allowing attackers to reach an endpoint via a certain path while providing a different value in the `.path`.\n\n### Impact\n- Instances of vllm that use an API Key for the OpenAI API and expose the API to attackers.\n- Instances behind an RFC-conforming web server (such as nginx) are **not** affected.",
  "id": "GHSA-94f4-hr76-p5j6",
  "modified": "2026-09-01T15:30:49Z",
  "published": "2026-06-16T17:36:41Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/vllm-project/vllm/security/advisories/GHSA-94f4-hr76-p5j6"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-48746"
    },
    {
      "type": "WEB",
      "url": "https://github.com/vllm-project/vllm/pull/43426"
    },
    {
      "type": "WEB",
      "url": "https://x41-dsec.de/lab/advisories/x41-2026-002-starlette"
    },
    {
      "type": "WEB",
      "url": "https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-48746.json"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/vllm-project/vllm"
    },
    {
      "type": "WEB",
      "url": "https://github.com/pypa/advisory-database/tree/main/vulns/vllm/PYSEC-2026-226.yaml"
    },
    {
      "type": "WEB",
      "url": "https://bugzilla.redhat.com/show_bug.cgi?id=2491581"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/security/cve/CVE-2026-48746"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:61629"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:61627"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:43038"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:42644"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:42142"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:42132"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:36006"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:36005"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:30089"
    },
    {
      "type": "WEB",
      "url": "https://access.redhat.com/errata/RHSA-2026:30088"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H",
      "type": "CVSS_V3"
    }
  ],
  "summary": "vLLM: OpenAI auth bypass"
}

No mitigation information available for this CWE.

No CAPEC attack patterns related to this CWE.